nonlocal bias

[changhoonhahn]

Test nonlocal bias model by comparing:

1. reduced bispectrum with local bias versus
2. reduced bispectrum with nonlocal bias with bestfit parameters alpha, beta from Francesco

[mcc1200]

all-k comparison of base ALPT+HOD vs nonlocal bias with alpha, beta, nmean pars fit by Francesco

[mcc1200]

nb to be posted

[changhoonhahn]

Reopening this issue so we can explore whether we need to include rhoeps or eps, in addition to alpha and `beta.

I explored the derivatives w.r.t. to the bias parameters:
https://github.com/changhoonhahn/hodalpt/blob/6d7a484e31465c655ee4fe1baf9781f8cafbe68d/nb/alpt_bias_priors.ipynb

A few conclusions based on this notebook:

• dth does not do much so let's 86 that
• Variations in the (knot, void), (filament, void), (sheet, void) do not contribute significantly so let's 86 them as well.
• rhoeps and eps seems to significantly impact clustering. Unclear whether it's strictly necessary.

[changhoonhahn]

A quick comparison of the power spectrum monopole predicted by a model w/ rhoeps and eps versus a model w/o reveals that we probably should include rhoeps and eps.

https://github.com/changhoonhahn/hodalpt/blob/main/nb/alpt_bias_extraparams.ipynb